Don’t look now, but AI is about to take an ax to a whole lot of traditional filmmaking procedures that we’ve been used to ever since the Lumiere brothers invented the moving picture.
For example, what if you could just source cast members from your computer instead of going out and putting out a call for auditions?
The work that this saves is going to be, well, expressed in orders of magnitude. Or in other words, everything’s about to change.
If you hear Joe Penna talking about this in his review of current technology and where it is, you get the picture pretty quickly – AI is here to stay. It’s something that everyone from the big-house runners to the stage hands is going to have to be thinking about.
As Penna talks about how he went from MysteryGuitarMan youtuber to director, and making his own movies, he also goes over the incredible progress made by technologies like Stable Diffusion in the last few months and years. Take a look at his recent presentation, and you’ll see a tale that, as someone who gets paid in the film world, is likely to make your hair stand on end.
It started with pictures.
Two years ago, we didn’t know how to make ourselves look like Van Gogh or Caravaggio paintings, or anything like that.
Now, nearly any kind of new illustrative visual can be made on the fly with Stable Diffusion or Dall-E.
In practical film applications, as Penna points out, you can work with a concept artist faster.
Not only that, but you can eliminate whole positions that used to be pretty labor-intensive.
When Penna talks about this, he goes over one of his projects as an example, where the producers needed a set of serial killer victims to show up in a book.
The problem, as you see from his testimony about what we’re doing with film now, is that nobody wanted to be paid “SAG minimum” to be a serial killer’s victim.
Not a problem – they just made some people up with AI, and stuck them right in the book!
When you talk to Penna, he also points out that this technology is going at lightning speed.
For example, where just a few months ago we had nonsense words up on a lot of AI visual images, people are now presumably working on running all of this stuff through a different neural net that will make sure it spells things right.
Read the full article here
